GPT-5 mini (medium)

August 7, 2025Last synced Apr 7, 2026, 4:00 PM

Blended Price

$0.69/M

Input Price

$0.25/M

Output Price

$2.00/M

Speed

85 tok/s

TTFT

18.67s

Benchmark Scores

Intelligence Index

External Benchmarks

ARC-AGI-1

How GPT-5 mini (medium) Compares

Axes

X Axis

Y Axis

Bubble Size

Filters

Blended Price (USD)$0.00 – $30.0

AgMoBench Overall≥ 3.5

Providers

ai21-labsalibabaanthropicawsazurebaiducoheredeepseek

Show quadrants

anthropic

openai

google

meta

mistral

nvidia

kimi

xai

azure

deepseek

aws

cohere

baidu

ai21-labs

zai

reka-ai

xiaomi

alibaba

minimax

ibm

Bubble size = Context Window

16384.00

2000000.00

Compare with other models

GPQA Diamond

0.8 / 100

LiveCodeBench

0.7 / 100

Terminal-Bench Hard

0.3 / 100

τ²-Bench

0.7 / 100

ARC-AGI-1 Cost per Taskarc_agi

0.0

ARC-AGI-2arc_agi

1.7 / 100

ARC-AGI-2 Cost per Taskarc_agi

0.0

AA-Omniscience AccuracyPredicted

40.3 / 100

AA-Omniscience Hallucination RatePredicted

95.9 / 100

Aider PolyglotPredicted

73.6 / 100

AIMEPredicted

0.6 / 30

AlpacaEval 2.0 LCPredicted

12.2 / 100

AlpacaEval 2.0 RawPredicted

9.5 / 100

BigCodeBench CompletePredicted

56.1 / 100

BigCodeBench InstructPredicted

45.5 / 100

AA Intelligence Index (Matrix)Predicted

68.3

AA Long Context Reasoning (Matrix)Predicted

73.1

AIME 2024Predicted

97.0

AIME 2025 (Matrix)Predicted

98.4

Arena-Hard AutoPredicted

82.8

BrowseCompPredicted

66.0

BRUMO 2025Predicted

96.0

CMIMC 2025Predicted

84.6

GPQA Diamond (Matrix)Predicted

85.6

HLE (Matrix)Predicted

25.2

HMMT Feb 2025Predicted

85.4

HMMT Nov 2025Predicted

77.3

HumanEvalPredicted

92.7

IFBench (Matrix)Predicted

54.0

IFEvalPredicted

90.3

IMO 2025Predicted

25.6

LiveCodeBench (Matrix)Predicted

78.7

MATH-500 (Matrix)Predicted

97.9

MathArena Apex 2025Predicted

1.9

MMLU-Pro (Matrix)Predicted

84.7

MMMU-ProPredicted

81.5

MRCR v2Predicted

79.0

OSWorldPredicted

42.0

SimpleQAPredicted

41.9

SMT 2025Predicted

87.5

SWE-bench ProPredicted

41.4

Tau-Bench Telecom (Matrix)Predicted

49.3

Terminal-Bench 2.0Predicted

33.7

Terminal-Bench 1.0Predicted

29.6

USAMO 2025Predicted

23.3

Video-MMUPredicted

86.5

browsecompPredicted

63.5

BullshitBenchPredicted

26.8 / 100

Aider PolyglotPredicted

0.1

Apex AgentsPredicted

2.5

BIG-Bench HardPredicted

3.0

CAD-EvalPredicted

3.2

Chess PuzzlesPredicted

0.2

CyBenchPredicted

0.4

DeepResearchBenchPredicted

0.4

GeoBenchPredicted

0.0

GSM8K (Epoch)Predicted

3.2

HellaSwagPredicted

0.1

Lech Mazur WritingPredicted

8.2

METR Time HorizonsPredicted

4.7

PosttrainbenchPredicted

0.0

The Agent CompanyPredicted

6.8

TriviaQAPredicted

8.7

WinoGrandePredicted

0.9

FrontierMathPredicted

19.3 / 100

GAIA Level 1Predicted

76.7

GAIA Level 2Predicted

63.8

GAIA Level 3Predicted

64.8

GAIAPredicted

61.3 / 100

LegalBenchPredicted

85.0 / 100

LongBench v2 EasyPredicted

72.9

LongBench v2 HardPredicted

47.1

LongBench v2Predicted

42.7 / 100

MATH-500Predicted

1.0 / 100

MathVistaPredicted

70.0 / 100

MedQA (USMLE)Predicted

91.7

MLE-benchPredicted

38.6 / 100

MMMUPredicted

80.4 / 100

MMTU Table UnderstandingPredicted

65.3 / 100

MT-BenchPredicted

7.9 / 10

NoLiMa (NIAH)Predicted

94.3 / 100

OCRBench v2Predicted

84.2 / 100

Open LLM AveragePredicted

11.9 / 100

Open LLM: BBHPredicted

39.3 / 100

Open LLM: GPQAPredicted

27.9 / 100

Open LLM: IFEvalPredicted

27.1 / 100

Open LLM: MATH Level 5Predicted

46.7 / 100

Open LLM: MMLU-PROPredicted

22.0 / 100

Open LLM: MUSRPredicted

38.6 / 100

RE-BenchPredicted

79.0

SimpleBenchPredicted

43.9 / 100

simpleqaPredicted

40.8

SWE-bench LitePredicted

45.1 / 100

tau-bench RetailPredicted

83.3 / 100

WebArenaPredicted

2.2 / 100

WildBenchPredicted

46.9

BFCL (Berkeley Function Calling)bfcl

55.5

Arc Agi 2epoch_ai

0.0

Epoch Capabilities Indexepoch_ai

145.0

FictionLiveBenchepoch_ai

0.6

OTIS Mock AIME 2024–2025epoch_ai

0.8

SimpleQA Verified (Epoch)epoch_ai

0.2

LiveBench Codinglivebench

68.2 / 100

LiveBench Data Analysislivebench

55.2 / 100

LiveBench Languagelivebench

75.5 / 100

LiveBench Mathlivebench

82.2 / 100

LiveBench Overalllivebench

65.9 / 100

LiveBench Reasoninglivebench

68.3 / 100

SWE-bench Verifiedswebench

56.2 / 100

Vectara Factual Consistencyvectara_hallucination

85.3 / 100

Vectara Hallucination Ratevectara_hallucination

14.7 / 100

WeirdMLweirdml

52.7 / 100